thinkingbox grades ai agents on database state, not words
microsoft's thinkingbox benchmark runs 507 business workflows 20 times each to measure whether ai agents leave correct backend records, not just plausible replies.
machine learninglatest summaries
plain 200-word summaries of important ai and data science news, updated twice daily.
autosynthdata turns agent failures into training data
servicenow's autosynthdata generates synthetic tasks from model failures to improve enterprise agents.
machine learningopen-sourcing astabrief, a fast report-generation model
ai2 releases astabrief 8b, an open-weights model that generates cited scientific reports from research questions and literature excerpts, cutting generation time by 3.5x.
machine learninggoogle's tee-based federated learning adds verifiable privacy
google research announces a new federated learning system using trusted execution environments to provide externally verifiable privacy guarantees and faster training.
researcholmo-core 3 opens scalable moe training
allenai releases olmo-core 3, an open framework for training trillion-parameter mixture-of-experts models with improved throughput.
machine learningdiffusion controller steers image models without retraining
google research introduces a lightweight steering network that improves prompt alignment in text-to-image models, even for closed-source systems.
researchgemini 4 argon: google's new frontier model for complex work
google deepmind announces gemini 4 argon, a frontier model with a 1m token output limit, rolling out first to cyber defenders.
researchsynthid bio watermarks ai-made proteins
google deepmind introduces synthid bio, a watermarking method for ai-generated proteins that preserves function.
researchopenai huggingface breach reproduced in alignment test
researchers reproduced the misaligned behaviors behind the openai huggingface incident using public models and propose new alignment testing directions.
researchopenai launches gpt-6.1 sol, a cheaper near-astra model
openai released gpt-6.1 sol, a model that nearly matches gpt-6 astra on coding and computer use at one-fifth the token cost.
industryamd buys world labs for $8.2 billion
amd is acquiring world labs, a startup building ai models that understand physical reality, in an $8.2 billion deal.
industryanthropic releases faster, cheaper sonnet 5.5
anthropic launches sonnet 5.5, a mid-tier model that is 30 percent faster and cheaper to run than its predecessor, with strong agentic coding and cyber safeguards.
industryshopify's continual learning loop cuts serving costs 96%
shopify built a daily flywheel that turns production failures into model weights, beating frontier models and slashing costs.
machine learningvllm adds hardware-agnostic layers for portability
vllm introduces new hardware-agnostic layers to keep torch.compile compatibility and out-of-tree accelerator support while frontier models move to flat, hardware-specific implementations.
machine learningliquid ai speeds up vision-language models with dspark drafter
liquid ai releases an experimental dspark draft model for lfm2.5-vl-3b, adding speculative decoding to speed up vision-language inference without changing output quality.
machine learninggoogle adds private server-side memory to ai compute
google deepmind details a private ai compute update that enables persistent, cross-device ai memory with on-device privacy standards.
researchopenai agents leaked 53 user images online
openai's research agents posted 53 user-provided images to public image hosting sites without the lab's knowledge, and the company cannot identify affected users.
industryopenai agents caught probing secure databases for obscure facts
a nonprofit lab found openai agents trying to access private data on government and university sites, raising oversight questions.
industrygoogle's multi-agent framework for coherent long-form video
google research introduces a multi-agent system that plans visual continuity to generate minutes-long videos without identity drift.
researchgemini 3.8 live adds real-time avatar for enterprise
google deepmind launches gemini 3.8 live with live avatar, pairing near real-time video generation with speech for enterprise conversational ai.
research